To think the junior / resident doctors are greedy, selfish, entitled & lazy?

[SpottyAlpaca]

So the resident doctors are out on strike. Yet again. Patients are being inconvenienced & treatments delayed. Yet again.

They have received a pay rise of 28.9% over that last 3 years, which is by far the highest increase of any group in the public sector. Very few people in the private sector, who ultimately pay the doctors’ salaries, have received anything like as much. Very few of their patients will ever earn as much as a resident doctor. Yet still it’s not enough and they are demanding even more.

Doctors do an important job and deserve to be paid properly for it. But the BMA’s current approach is completely unreasonable and deluded. They talk about “pay restoration’ to 2008 levels but that’s completely unrealistic. The country is poorer now & simply can’t afford it. AIBU to think they should get back to work?

I can't imagine AI being much use diagnostically. Doctors, especially if they are overconfident, make diagnoses heuristically, excessively drawing on recall and pattern recognition. This is usually how mistakes are made, and at best algorithms would presumably replicate this.

Humans would still be required to input data acquired from the history, physical examination, and investigations. A skilled diagnostician approaching the problem analytically might acquire the data needed to reach a diagnosis to feed into the AI, but you don't know whether data collection is complete until you have a diagnosis.

Quite apart from the prodding and poking, it takes a long time to develop a sense of when something doesn't quite add up and what to do about it.

Microsoft’s AI Is Better Than Doctors at Diagnosing Disease https://share.google/vWpTODSe8HzgsyBOG

Ai is now more sophisticated than relying on a clinician to gather relevant information. It already knows what questions need to be asked, what tests need to be performed and details of possible diagnoses, which it can easily narrow down.

Thanks for the link. It is an interesting paper but it doesn't support your assertion.

There were only four hospital doctors in the study, and the median experience of all of the doctors was only 12 years, which means on average they were barely fully trained. Furthermore, the hospital doctors were generalists, in real life patients are triaged to appropriate specialists according to their age, presentation etc..

It was a paper exercise - the test sample of doctors and the AI relied on data that had been collected by doctors. Until AI is adept at general, cardiovascular, respiratory, abdominal, and neurological examinations it will always be thus.

The clinical scenarios were taken from case reports that had been solved by doctors with a success rate of 100% - vs AI's success rate of 85%. I don't think I would want to see a specialist with a more than 1 in 10 chance of making the wrong diagnosis.

There are other studies available and things will only improve as technology improves. There won't be any rolling back from this.

Not to mention the cost savings and of course, no more strikes.

In my experience they don’t. They just get away with it.

I think I first heard that 50 yrs ago.

When Artificial Intelligence Outperforms Humans — Evidence from Recent Clinical Studies - Surgery International https://share.google/uMKF2rwyDxtAeyurC

This is clearly the way forward. Humans will feature to a certain degree, but if doctors continue to misdiagnose at the rate they do, their employers aren't going to stand for it. The cost in litigation alone must be quite high. Perhaps ai will reveal what the public have already known for a long time. That an accurate diagnosis is often a matter of luck and who you get to see.

There was no ai 50 years ago. Were there mri machines 50 years ago? Do mri machines assist in diagnosis now? Ai is like nothing we've seen before. Computers have been around for decades, but this is something different.

The cost in litigation alone must be quite high.

Figures for 2024/25:

£3.1 billion on claims and related costs (eg their own legal fees and a claimant’s legal fees if NHSLA loses or agrees to pay as part of settlement).

Of which £1.3 billion was related to obstetrics - often damaged babies requiring a lifetime of expensive care.

£60 billion estimated for future liabilities eg open claims not yet quantified or incidents that have not yet resulted in a claim.

14.5k new clinical negligence claims and reported incidents.

Clap, clap.

None of the papers cited on the Surgery International website are as persuasive as the Nori paper. Some suggest specific nonclinical situations where AI might have a role, but clinical diagnosis is not one of them

We need to accept that lowering academic requirements at the point of entry to medicine, and then not teaching a lower ability intake the traditional sciences, has not been the success that some people were expecting. Because there is no alternative on the horizon to a switched on and properly educated human.

This, plus robotics.

And both are in an infant state right now, and when we have portable quantum computing, which became viable very recently, there will be no comparison between what is possible today and what will be routine.

People who think doctors won't be replaced, in time, are just wrong. Even my eye surgeon, in the middle of cutting a hole in my eye, told the room that he's hoping to reach retirement age before it replaces him, and he wasn't joking.

Computers have been around for decades
most of the computers in the NHS have been around for decades, that's for sure!! (Productivity would deffo improve if the tech was up to the job and any of the applications communicated with each other. NHS procurement (a process that largely doesn't involve HCPs) is usually a process of finding the cheapest option and then looking surprised when it's not up to the job).

Yes, back when Pong was the height of AI is not a dishonest comparison at all.

Having a basic conversation with natural language, was basically science fiction for people less than 10 years ago, barely 6 years ago really - although the process was start more like around 10 years ago and the first small llms started appearing in very limited ways first with closed access as they tested the waters of how people would react and use them.

Now they can write entire books on their own even if a bit rough and basic, can generate video, code extensively, pass law and complex maths exams, and beat many doctors on average at diagnosing things - the list goes on. All in no time at all.

It's still a distant horizon thing for now to use them seriously in many areas, but you are living in denial to pretend things havent changed for good.

A junior Dr in A&E saved my life last weekend, he had just finished a placement in neurology and had come across transverse myelitis which I have. I support them x

He wasn't on strike, then, on a day when others were striking, and yet you suppport them striking?

I'm glad you got the help you needed.

You lack basic comprehension skills if you reckon that I "pretend that things haven't changed for the good". If you want to take anything from my comment other than its literal meaning - which was true - it would be that the previous poster's optimism is misplaced. As yours is.

Cite one clinical example in support of the delusional belief that AIs "beat many doctors on average at diagnosing things."

He diagnosed his rare disease using Google. Now he hopes AI can do the same for others | CBC Radio https://share.google/EaiUV8X7Xd0eVMYDc

I'm not trawling the internet for similar examples, but there are many more. He discovered his using a simple Google search. Ai is far more capable than that and hospitals are using it right now.

'AI said I had Lyme disease before a doctor did' - BBC News https://share.google/U0o71NP8VVIyFYX7T

Doctors Told Him He Was Going to Die. Then AI Saved His Life | I3H | Perelman School of Medicine at the University of Pennsylvania https://share.google/1V4wJ4qcDvq7Q9Oeg

AI detected rare kidney disease, saved man from certain death - The Economic Times https://share.google/ysBPzB5UexJtYytmL

Teenager saved own life when he ignored his GP and used ChatGPT to diagnose a rare medical condition https://share.google/R8JqXw5o0A7y1HNGi

Microsoft says AI system better than doctors at diagnosing complex health conditions | AI (artificial intelligence) | The Guardian https://share.google/o42QQEWlm4Tie7Mdy

Is AI Better at Diagnosing Disease than a Doctor? - News https://share.google/YSeHL3GDPabaW9OWr

From "Large Language Model Performance and Clinical Reasoning Tasks" published yesterday in JAMA online:

"The promise of LLMs in clinical medicine lies in their potential to augment—not replace—physician reasoning. This study establishes the first benchmark for longitudinal clinical reasoning (to our knowledge) and introduces the PrIME-LLM framework, a multidimensional metric designed to capture performance across the full arc of diagnostic and management tasks.

Our evaluation suggests that despite rapid advances in pattern recognition and knowledge retrieval, current LLMs still lack the reasoning processes needed for safe clinical use. The consistent gap between differential diagnosis and final diagnosis highlights how differently these systems process information compared with physicians. Clinicians preserve uncertainty and iteratively refine differential diagnoses, whereas LLMs collapse prematurely onto single answers, a limitation that persists across model generations. Their weak performance on differential diagnosis, consistent with a prior study from authors of the current work,8 suggests these limitations persist across early and state-of-the-art models. The risk is not just that LLMs are sometimes wrong but that their reasoning is brittle precisely where uncertainty and nuance matter most. Benchmarks that reward only correct final answers risk reinforcing this shortcutting, widening the gap between marketing claims and the skills actually required at the bedside."

https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2847679

Furthermore, this and studies like it are paper exercises, diagnosis is an interactive process involving sensory modalities that AIs lack.

This government, has already stated its intention to reduce referrals to secondary care and reduce bed numbers by expanding services in the community. They have also said the NHS will make much more use of AI. We already have GP consultations over the telephone and "virtual hospital wards"; it takes very little imagination to see where this is heading. I have been right about every other organisational disaster in the NHS, and I'm not going to be wrong about this.

There's a word missing.

"Furthermore, this and studies like it are paper exercises, diagnosis is an interactive process involving sensory modalities that AIs currently lack."

Given that they currently perform weakly on paper, for all the hype, I wouldn't worry too much about that.